Emotional power of live music
Sascha Frühholz describes the emotional power of live music compared to recorded music.

PNAS: Welcome to Science Sessions, the podcast of the Proceedings of the National Academy of Sciences, where we connect you with Academy members, researchers, and policymakers. Join us as we explore the stories behind the science. I’m Paul Gabrielsen.
Why does live music have a different emotional power than recorded music? Is it in the ability of a musician to adapt their performance to the emotions of the audience? In a recent PNAS study, Sascha Frühholz, now at the University of Oslo, and colleagues at the University of Zurich compared brain activity during live and recorded music performances. The live performers were able to adapt their music to elicit an enhanced emotional response from listeners. The live music produced a stronger and more consistent brain activity response than recorded music and produced a synchrony between performer and listener that recorded music couldn’t duplicate.
Sascha, what did we know before your study about the neuroscience of emotional responses to music?
Frühholz: Music is powerful in terms of portraying emotion. It’s also powerful in eliciting emotion in listeners. This is also what all the previous studies have shown using recorded music. If you enjoy listening to music, we see a lot of activity in the pleasure center of the brain. It can stimulate the auditory system if you like the acoustic quality of the music. And it can also stimulate what we call the affective brain system, which we also call the limbic system, which is actually responding to the emotion quality, if it’s positive music, if it’s negative music.
Unfortunately, and this was also a little bit the motivation for the study, previous research didn’t show consistent activity, especially in this limbic system. So there were some studies [that] reported, “oh we find activity in the amygdala indicating that people emotionally responded to the music.” But, surprisingly, there were other studies that didn’t show any kind of activity in the amygdala, and we thought, “OK, why is this the case?” So we thought we’d go for live music.
PNAS: Why did you hypothesize that live music would show a different neural response?
Frühholz: If you go to a concert, I mean, the emotion intensity is so much higher than listening to recorded music. So this was one of the ideas, just using this intensified emotional quality of live music, which we thought then would also stimulate the brain much, much better than recorded music. If you perform as a musician, you want to perform to increase, really, the emotional response in the audience. So if musicians perform, they see the response of the audience, and then they can adapt their performance.
PNAS: Tell me about your experimental setup.
Frühholz: Actually, I have to tell you, it was quite difficult! In the end, it worked. We had piano players in one room, and just in the room next to the piano players, we had people inside what we call an MR scanner—that’s a big machine that is recording brain activity.
And we were recording brain activity in real time—especially in the system that we call the limbic system, especially in the amygdala. And the musicians in the other room, they could live follow the brain activity of the person inside the scanner while listening to the performance. And the task for the piano players was: If you see that the signal in the amygdala is going down, if it’s decreasing, try to change something in your performance in order to increase this activity again.
PNAS: Did the performers play the same pieces of music for each listener?
Frühholz: It was the same set of pieces for each listener. So we composed short piano pieces before the experiment. We recorded these piano pieces also before the experiment. But the same piano players play the same pieces live during the experiment.
[Music plays]
The pieces [themselves], they were composed in a sense of what we call jazz standards. It allows a lot of improvisation while performing the musical pieces.
We instructed the participants to try to introduce modulation in three different modalities, or three different dimensions. One is what we call the so-called density of the events, which actually means, if you play piano, you can decide, “how many keys do I play on the keyboard?”
[Example of density modulation plays]
The more keys you play on the keyboard, the more complex the piano song becomes. One other type of modulation was the so-called dynamic modulation. So if you play piano, you can decide how hard I strike the key.
[Example of dynamic modulation plays]
And the third one was what we call articulation. It actually means how much you emphasize a specific theme or topic in the piano pieces.
[Example of articulation modulation plays]
PNAS: What did you find? How did your participants respond to the music?
Frühholz: When musicians played the pieces in the live condition, it could more strongly stimulate the brain. It could also more consistently stimulate the brain, especially, again, this part which we call the affective brain system. It also stimulated many other different brain areas. We also observed that many of these regions connected more with each other during the live condition. During live music, there are so many things happening, so many changes, and you have to process these changes and you need a lot of brain areas. So this is also why all these regions actually connect.
PNAS: What could live music do that recorded music couldn’t?
Frühholz: One thing that we also quantified is what we call synchrony between the musician and the listener. We quantified when piano players were changing something in the performance, and we also then quantified: Is there an immediate change in how the brain of the listener is actually responding? And if you see a similar change on both ends, so in the performance and in in the listener—this actually happened during the live music, we had this synchrony between musicians and the listeners. This is what we didn’t observe during the recorded conditions. So it was the same piece, same listeners, but there was not a synchrony between the musical piece and the brain of the listeners.
PNAS: Did the listeners know that they were listening to live or recorded music?
Frühholz: Not before the experiment. We just told them they were going to [be] listening to music. But they could not really say, “This was a live piece,” “This was a recorded piece,” but they could recognize there was a difference between the live and the recorded condition. Once we explained [it to] them, then, they just suddenly realized, “yes there was a live condition, there was a recorded condition in the experiment.”
PNAS: There was some unpleasant music included as well. Why did you include that? What did it show?
Frühholz: So we included what we call unpleasant music.
[Example of unpleasant music plays]
Unpleasant music is sometimes dissonant, is often in minor mode, could evoke feelings, let’s say, of sadness but also, like, feelings of aversion. We included positive music, so mainly, let’s say, pieces in major mode, happy emotions…
[Example of pleasant music plays]
…simply because we want to see: Is the brain responding differently to positive emotion, expressions, music and to negative emotions? What we found: It works both, so you can stimulate the affective brain system both with pleasant music and with unpleasant music.
PNAS: How did the listeners feel about the experience?
Frühholz: Sometimes they could not put it in words, but they could really see it, when we told them there was a live condition and a recorded condition. Sometimes they said the music during the live condition was more engaging. They much more liked listening to this music, and they didn’t say the words “It was live music” or “It was recorded music,” they just described it in qualitative terms actually.
PNAS: What did the musicians think of participating in this experiment?
Frühholz: For them, in the beginning, it was some type of trial and error. We just gave them the task: Try to change something during the performance. But they didn’t know really what to change. They tried different things, and I would say they really quickly then realized “OK, for this listener specifically, I have to change something like the tempo, for example; I have to change the timbre [of] the music.” So was quite an experience also for the musicians in the end.
PNAS: How generalizable are these findings to other types of music, people, settings, or situations?
Frühholz: I would say you can generalize them to a certain degree. What we simulate in our experiment is a live setting where musicians really can change their performance. That’s not always possible during every type of concert. Classical concerts, for example, it’s quite difficult for the orchestra to respond to the audience. But I would say like some small club concerts where you have singer-songwriters, jazz concerts, then it’s much more possible because then also the musicians are much more in connection with the audience.
PNAS: So, this is a small study with a small sample of participants, yet it lends interesting insights for further exploration. What are the caveats or limitations of this study?
Frühholz: There was only one musician and there was one person listening, and, usually, if you go to a live concert, it’s not only one person listening; you really have a lot of people in the audience. Live concerts—the intensified emotion not only comes from the performance of the musician, it’s also, like, the social context. If I feel an emotion, I see other people in the audience also feel the same emotion; it amplifies itself on a social level.
One limitation is we had piano music. It was a single piano player. It was not a band playing, it was a single musician. In live concerts if you have, like, a band on stage; it’s not only one musician responding to the audience; it’s really all people on stage are responding.
PNAS: Thanks for tuning into Science Sessions. You can subscribe to Science Sessions on iTunes, Spotify, or wherever you get your podcasts. If you liked this episode, please consider leaving a review and helping us spread the word.
All podcast episodes are published with CC BY-NC-ND license.